Poor prognosis cancers of the lung, esophagus, pancreas, and liver account for 40% of all cancer deaths and a sizable share of overall cancer care costs. Despite little improvement in survival rates over time, how intensively patients are treated has increased dramatically. Treatment intensity, as reflected by average Medicare expenditures, varies widely across health systems. The root causes of such variation and its implications for patients have not been assessed in a systematic fashion. To this end, the candidate (Dr. Sandra L. Wong) seeks to examine treatment intensity in order to gain a better understanding of variation in cancer care and its implications. The overall goal of this application is to support the candidate's continued training and development as an independent investigator in cancer-focused, policy-relevant health services research. The career development plan is based on coursework, guidance from mentors, and the practical application of skills through research. Training goals for the period supported by this award mechanism include acquiring expertise in advanced quantitative methods, acquiring expertise in healthcare systems and policy issues, and advancing career development skills. The candidate's main research goals are to better understand variation in the care of patients with cancer. The major themes of the candidate's research interests are reflected in the Specific Aims of this proposal: (1) to characterize variation in the delivery of health care at the health system level, (2) to identify determinants of variation, and (3) to understand the relationship between variations and outcomes. National Medicare data will be used to identify which types of services account for variation in overall expenditures, including the use of diagnostic studies, surgical procedures, systemic chemotherapy, radiation therapy, and end-of-life care. The main determinants of variation will then be examined, focusing on both patient factors (such as race and socioeconomic status) and systems-level factors. Finally, the relationship between treatment intensity and patient outcomes will be assessed. This project will lay a foundation for identifying areas of both overtreatment and under-treatment of patients with poor prognosis cancers and facilitate efforts to improve both efficiency and quality of care. PUBLIC HEALTH RELEVANCE: For patients with a poor prognosis cancer, their treatment depends on where they live and in which health system they receive their care. Focusing on the national Medicare population, the proposed work will (1) examine variation in the intensity of care for poor prognosis cancers, (2) identify patient- and systems-level determinants of variation in treatment, and (3) assess relationships between intensity of care and patient outcomes. Broadly, findings will lay a foundation for identifying areas of both overtreatment and under-treatment and facilitate efforts to improve efficiency and quality of care.